INDEX

Explanations

words related to receiving awards or being recognized for achievement

oai_token-act-pair · gpt-3.5-turbo

mentions of various types of awards

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.2.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

loo

-0.86

 Pastebin

-0.74

Alz

-0.70

 Stur

-0.65

sei

-0.64

 Witt

-0.63

STD

-0.63

olen

-0.61

Geo

-0.60

alter

-0.60

POSITIVE LOGITS

 awards

1.24

 award

1.16

 laure

1.09

 awarded

1.06

 prizes

0.99

 awarding

0.93

 accol

0.92

Winner

0.89

 prize

0.88

 Award

0.88

Activations Density 0.013%

words related to receiving awards or being recognized for achievement

mentions of various types of awards

No Comments

No Known Activations